owner stringclasses 14
values | repo stringclasses 14
values | id int64 116k 4.61B | issue_number int32 1 181k | author stringlengths 1 39 | body stringlengths 1 262k | created_at timestamp[us]date 2000-06-06 02:40:44 2026-06-02 20:16:18 | updated_at timestamp[us]date 2000-06-06 02:40:44 2026-06-02 20:22:56 | reactions unknown | author_association stringclasses 5
values |
|---|---|---|---|---|---|---|---|---|---|
ClickHouse | ClickHouse | 715,625,697 | 16,295 | alexey-milovidov | > This task is already in the Intern List.
No, the task in intern's list is about schema inference from the data - regardless to the existence of INSERT query.
And what's proposed in this task is to:
- figure out that we have INSERT SELECT;
- make sure SELECT is trivial `SELECT * FROM function(...)`;
- provide... | 2020-10-23T23:04:52 | 2020-10-23T23:04:52 | {} | MEMBER |
ClickHouse | ClickHouse | 715,631,996 | 16,247 | vzakaznikov | > @vzakaznikov Look at the Testflows check.
Should be fixed now. | 2020-10-23T23:32:23 | 2020-10-23T23:32:23 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 715,668,555 | 16,330 | alexey-milovidov | The storage on disk is nearly optimal, because it is using LZ4 that acts like RLE in degenerate case - already compresses runs of consecutive values efficiently. It is not optimal if random distribution of NULLs is present, though.
The performance drop is associated with the need to check for NULLs (see `AggregateFu... | 2020-10-24T03:53:50 | 2020-10-24T03:53:50 | {} | MEMBER |
ClickHouse | ClickHouse | 716,077,055 | 16,343 | alexey-milovidov | Yes, it helped.
PS. The queries with `and` and `or` are unstable - obviously due to large footprint of data in cache (64K * 11). | 2020-10-25T01:04:16 | 2020-10-25T01:04:16 | {} | MEMBER |
ClickHouse | ClickHouse | 716,086,016 | 16,236 | trollhe | > Parquet format is very memory intensive for clickhouse to work with currently. You can perhaps lower the concurrency if it's some sort of ETL query: `settings max_threads = 1`
thanks amosbird, i setting `max_threads = 1`,but new error to me.bellow, see my command detail:
clickhouse-client
```
SET max_memory_u... | 2020-10-25T03:00:04 | 2020-10-25T03:00:04 | {} | NONE |
ClickHouse | ClickHouse | 716,137,145 | 15,983 | alesapin | Performance looks OK. | 2020-10-25T12:06:02 | 2020-10-25T12:06:02 | {} | MEMBER |
ClickHouse | ClickHouse | 715,385,062 | 16,251 | den-crane | No. The limit is also in GB(1000). Check my message
max_partition_size_to_drop = 5000000000
1. Size (10.21 GB) is greater than max_[table/partition]_size_to_drop (5.00 GB) | 2020-10-23T14:42:11 | 2020-10-23T14:42:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 715,405,053 | 16,149 | abyss7 | Yandex third-party fixed | 2020-10-23T15:15:29 | 2020-10-23T15:15:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 715,562,689 | 15,428 | haoch | This PR included some unrelated changes, recreate another PR with clean code at https://github.com/ClickHouse/ClickHouse/pull/16309 cc: @amosbird | 2020-10-23T19:59:42 | 2020-10-23T19:59:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 715,845,639 | 12,632 | jsding | I found
> Once a table is created, its storage policy cannot be changed
https://clickhouse.tech/docs/en/engines/table-engines/mergetree-family/mergetree/
So there is no way to change the storage policy for existing tables. :( | 2020-10-24T07:18:16 | 2020-10-24T07:18:16 | {} | NONE |
ClickHouse | ClickHouse | 716,070,879 | 16,345 | den-crane | No. Not in the near future. | 2020-10-25T00:00:30 | 2020-10-25T00:01:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 716,096,585 | 16,347 | alexey-milovidov | Testflows check - bad test, @vzakaznikov | 2020-10-25T05:23:33 | 2020-10-25T05:23:33 | {} | MEMBER |
ClickHouse | ClickHouse | 715,382,306 | 16,251 | keliss | But even in this case the exception is triggered by a lower amount of disk space than the limit :)
48.43 GiB * 1024 / 1000 = 49.59 GB. | 2020-10-23T14:37:31 | 2020-10-23T14:37:42 | {} | NONE |
ClickHouse | ClickHouse | 715,661,814 | 16,008 | zhang2014 | Fixes: https://github.com/ClickHouse/ClickHouse/issues/15883#issuecomment-708932203 and https://github.com/ClickHouse/ClickHouse/issues/15883#issuecomment-708939294 | 2020-10-24T02:52:25 | 2020-10-24T02:52:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 715,665,100 | 16,285 | sundy-li | BTW, the latest croaring libs already have 3 kinds of `bitmap containers` for different kinds of data.
- BITSET container (for large data)
- ARRAY container (for small data)
- RUN container (RLE compress)
| 2020-10-24T03:19:53 | 2020-10-24T03:19:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 715,694,183 | 16,333 | alexey-milovidov | > /ClickHouse/docker/test/fasttest/run.sh: line 175: COPY_CLICKHOUSE_BINARY_TO_OUTPUT: unbound variable | 2020-10-24T04:59:35 | 2020-10-24T04:59:35 | {} | MEMBER |
ClickHouse | ClickHouse | 715,695,583 | 16,331 | alexey-milovidov | Bad error message - it's for the arguments of file-like table functions. | 2020-10-24T05:01:07 | 2020-10-24T05:01:07 | {} | MEMBER |
ClickHouse | ClickHouse | 715,878,588 | 16,332 | alexey-milovidov | Testflows - bad tests, @vzakaznikov | 2020-10-24T08:11:17 | 2020-10-24T08:11:17 | {} | MEMBER |
ClickHouse | ClickHouse | 715,922,387 | 16,249 | vzakaznikov | The problem is that we can't reliably detect when LDAP external user directory is ready to authenticate users.
| 2020-10-24T14:24:28 | 2020-10-24T14:24:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 716,071,329 | 16,345 | den-crane | https://youtu.be/7pK-rrmnt8Q?t=6179 | 2020-10-25T00:04:35 | 2020-10-25T00:04:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 715,548,144 | 16,261 | tavplubix | > `01516_drop_table_stress` -- timeout https://clickhouse-test-reports.s3.yandex.net/16261/b29cfa7f826566192478c77413ccca34e8a719ac/functional_stateless_tests_(memory).html#fail1
Funny race between async `DROP DATABASE` and `DROP TABLE SYNC`, fixed in #16127 | 2020-10-23T19:39:01 | 2020-10-23T19:39:01 | {} | MEMBER |
ClickHouse | ClickHouse | 716,085,147 | 16,281 | alexey-milovidov | @trollhe It's better to put requests in a queue and execute sequentially (or limit parallelism to a relatively small value) than to do all 2000 queries at once. | 2020-10-25T02:47:55 | 2020-10-25T02:47:55 | {} | MEMBER |
ClickHouse | ClickHouse | 715,338,579 | 16,242 | akuzm | Apparently the SQL-standard syntax for this is `select tuple.*`:
https://www.postgresql.org/docs/current/rowtypes.html#ROWTYPES-USAGE | 2020-10-23T13:24:19 | 2020-10-23T13:24:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 715,403,565 | 5,371 | yehiaelbehery | @sundy-li It would be awesome if you can do it. Thanks. | 2020-10-23T15:13:04 | 2020-10-23T15:13:04 | {} | NONE |
ClickHouse | ClickHouse | 715,465,815 | 16,300 | amosbird | I'm also investigating similar issues, and it's better to have JBOD balancing at partition-level. | 2020-10-23T17:09:44 | 2020-10-23T17:09:44 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 715,467,836 | 16,295 | amosbird | This task is already in the Intern List. However I do have another simple implementation in mind. We can provide a `schema` function which accepts a table identifier and evaluates to a const string of column definition list. We can also provide some abilities to transform the result list. | 2020-10-23T17:13:59 | 2020-10-23T17:13:59 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 715,661,881 | 16,008 | zhang2014 | CC: @tavplubix | 2020-10-24T02:53:01 | 2020-10-24T02:53:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 716,072,173 | 16,344 | alexey-milovidov | Testflows - bad tests, @vzakaznikov | 2020-10-25T00:13:35 | 2020-10-25T00:13:35 | {} | MEMBER |
ClickHouse | ClickHouse | 715,387,282 | 16,251 | den-crane | and
`48.43 * (1024 * 1024 * 1024) / (1000*1000*1000) = 52.00`
>1. Size (52.01 GB) is greater than max_[table/partition]_size_to_drop (50.00 GB) | 2020-10-23T14:46:14 | 2020-10-23T14:46:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 715,468,189 | 16,199 | azat | @alexey-milovidov so is it OK for upstream? | 2020-10-23T17:14:45 | 2020-10-23T17:14:45 | {} | MEMBER |
ClickHouse | ClickHouse | 715,654,667 | 16,328 | alexey-milovidov | It's the effect of enabled ThinLTO after moving to clang. There is also performance improvement as a bonus. | 2020-10-24T01:50:16 | 2020-10-24T01:50:16 | {} | MEMBER |
ClickHouse | ClickHouse | 715,671,797 | 16,269 | alexey-milovidov | Minor fix. | 2020-10-24T04:31:23 | 2020-10-24T04:31:23 | {} | MEMBER |
ClickHouse | ClickHouse | 715,938,385 | 16,281 | den-crane | There are a limiter `max_concurrent_queries=100`.
https://clickhouse.tech/docs/en/operations/server-configuration-parameters/settings/#max-concurrent-queries
It limits the number simultaneously processed requests.
How many queries you execute simultaneously? | 2020-10-24T15:31:18 | 2020-10-24T15:31:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 716,034,081 | 16,249 | alexey-milovidov | @vzakaznikov Let's make a loop with retries, about one retry per second and up to 100 retries in total. | 2020-10-24T18:16:42 | 2020-10-24T18:16:42 | {} | MEMBER |
ClickHouse | ClickHouse | 716,100,205 | 16,347 | alexey-milovidov | > Performance — 7 faster | 2020-10-25T06:14:01 | 2020-10-25T06:14:01 | {} | MEMBER |
ClickHouse | ClickHouse | 715,339,820 | 16,242 | akuzm | > Apparently the SQL-standard syntax for this is `select tuple.*`:
>
> https://www.postgresql.org/docs/current/rowtypes.html#ROWTYPES-USAGE
So maybe it should be in `ColumnsTransformer` or how is it called, the same place that rewrites `*` into the column list? | 2020-10-23T13:26:41 | 2020-10-23T13:26:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 715,339,838 | 16,242 | amosbird | > Apparently the SQL-standard syntax for this is `select tuple.*`:
>
> https://www.postgresql.org/docs/current/rowtypes.html#ROWTYPES-USAGE
But that's not the same. It's more like Asterisk vs Qualified Asterisk, and the latter introduces names that are not easy to refer to. | 2020-10-23T13:26:43 | 2020-10-23T13:28:39 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 715,398,739 | 16,284 | amosbird | > > Perhaps we can serialize the metadata with `ORDER BY` clause generated?
>
> Yes, but it can be surprising for the user. You just typed one `CREATE` query, and `SHOW CREATE TABLE` shows you different one.
Currently it already shows different things (such as settings index_granularity = 8192) | 2020-10-23T15:04:56 | 2020-10-23T15:04:56 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 715,449,792 | 16,251 | keliss | Ok, got it. Thanks for your help. | 2020-10-23T16:37:03 | 2020-10-23T16:37:03 | {} | NONE |
ClickHouse | ClickHouse | 715,562,494 | 16,309 | haoch | This PR is to resolve #16108 | 2020-10-23T19:59:16 | 2020-10-23T19:59:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 715,610,852 | 16,295 | alex-zaitsev | I think it is different/simpler than a task in interns list (https://github.com/ClickHouse/ClickHouse/issues/14450). When insert/select is considered the destination structure is already given, so source structure should match. clickhouse-client can already do it when accepting insert from stdin. | 2020-10-23T22:06:00 | 2020-10-23T22:06:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 715,340,510 | 16,251 | keliss | No, this partition remained intact for sure, we don't insert data for some past period of time. | 2020-10-23T13:27:56 | 2020-10-23T13:27:56 | {} | NONE |
ClickHouse | ClickHouse | 715,359,497 | 16,284 | amosbird | Perhaps we can serialize the metadata with `ORDER BY` clause generated? | 2020-10-23T13:58:44 | 2020-10-23T13:58:44 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 715,405,277 | 16,261 | alesapin | Waiting for normal stateless tests | 2020-10-23T15:15:57 | 2020-10-23T15:15:57 | {} | MEMBER |
ClickHouse | ClickHouse | 715,661,866 | 16,072 | alexey-milovidov | I just get tired of reports that something does not build on some version of AppleClang.
For Linux with clang and gcc, we have CI and we ensure that everything will go smooth.
Also we have cross-compiled builds that produce official binaries for Mac OS.
Do you know any reason why it builds successfully on your mac... | 2020-10-24T02:52:52 | 2020-10-24T02:54:02 | {} | MEMBER |
ClickHouse | ClickHouse | 715,671,790 | 16,272 | alexey-milovidov | Minor fix. | 2020-10-24T04:31:17 | 2020-10-24T04:31:17 | {} | MEMBER |
ClickHouse | ClickHouse | 716,063,228 | 14,840 | alexey-milovidov | Duplicate of #6420 | 2020-10-24T22:24:54 | 2020-10-24T22:24:54 | {} | MEMBER |
ClickHouse | ClickHouse | 716,083,698 | 16,281 | trollhe | > There are a limiter `max_concurrent_queries=100`.
> https://clickhouse.tech/docs/en/operations/server-configuration-parameters/settings/#max-concurrent-queries
> It limits the number simultaneously processed requests.
> How many queries you execute simultaneously?
thank you . i need to execute about 2000 querie... | 2020-10-25T02:29:01 | 2020-10-25T02:29:01 | {} | NONE |
ClickHouse | ClickHouse | 716,116,501 | 16,008 | tavplubix | AST fuzzer - #16081
Performance - unrelated | 2020-10-25T09:13:02 | 2020-10-25T09:13:02 | {} | MEMBER |
ClickHouse | ClickHouse | 715,367,925 | 16,251 | den-crane | Ah, it is two different units Gib vs GB 1024 vs 1000.
```
<max_partition_size_to_drop>5000000000</max_partition_size_to_drop>
SELECT
formatReadableSize(sum(bytes_on_disk)),
round(((sum(bytes_on_disk) / 1000) / 1000) / 1000, 2) AS GB
FROM system.parts
WHERE table = 'XX'
┌─formatReadableSize(su... | 2020-10-23T14:12:50 | 2020-10-23T14:13:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 715,446,396 | 16,284 | alesapin | > > > Perhaps we can serialize the metadata with `ORDER BY` clause generated?
> >
> >
> > Yes, but it can be surprising for the user. You just typed one `CREATE` query, and `SHOW CREATE TABLE` shows you different one.
>
> Currently it already shows different things (such as settings index_granularity = 8192)
... | 2020-10-23T16:30:30 | 2020-10-23T16:30:30 | {} | MEMBER |
ClickHouse | ClickHouse | 715,669,131 | 16,331 | alexey-milovidov | There is no such thing as `Content-Encoding: auto` in HTTP. | 2020-10-24T04:00:02 | 2020-10-24T04:00:02 | {} | MEMBER |
ClickHouse | ClickHouse | 716,136,871 | 16,265 | liangli1213 | @zhang2014 , as I explained, there is some sensitive data in bin log so I can't provide it to you.
And I have written a script to do sanity check between MySQL data and ClickHouse data, and I found the issue only happenes for decimail column, and with very low probability, here is a result for one of the tables:
``... | 2020-10-25T12:03:52 | 2020-10-25T12:03:52 | {} | NONE |
ClickHouse | ClickHouse | 715,380,166 | 16,265 | liangli1213 | @BohuTANG , I saw MySQL BinLog file check tool is available in 20.11.1.4978, so what is the next step?
Also I have some questions:
1. I have checked the MySQL side data, it is '0.3333', do you think the according bin log event record wrong information?
2. As you see the record, do you think the wrong one also has ... | 2020-10-23T14:33:45 | 2020-10-23T14:33:45 | {} | NONE |
ClickHouse | ClickHouse | 715,405,109 | 16,261 | alesapin | 
| 2020-10-23T15:15:37 | 2020-10-23T15:15:37 | {} | MEMBER |
ClickHouse | ClickHouse | 715,663,324 | 15,364 | alexey-milovidov | It's very unusual application :)
The downside is that the whole file will be buffered in memory. | 2020-10-24T03:04:16 | 2020-10-24T03:04:16 | {} | MEMBER |
ClickHouse | ClickHouse | 715,669,656 | 10,540 | ucasfl | > * [ ] 1. Implement `UNION DISTINCT`. For implementation just add Distinct transform after union.
> Add syntax `UNION` without ALL or DISTINCT and setting `union_default_mode` (similar to `join_default_strictness`) with values: empty string, ALL, DISTINCT. When set to empty string, UNION without ALL or DISTINCT wi... | 2020-10-24T04:05:36 | 2020-10-24T04:05:36 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 715,670,957 | 12,632 | jsding | Hi, Any comments on changing existing table's storage policy? No document found.
Thanks for your help | 2020-10-24T04:21:46 | 2020-10-24T04:21:46 | {} | NONE |
ClickHouse | ClickHouse | 715,671,780 | 16,271 | alexey-milovidov | Minor fix. | 2020-10-24T04:31:14 | 2020-10-24T04:31:14 | {} | MEMBER |
ClickHouse | ClickHouse | 715,671,793 | 16,270 | alexey-milovidov | Minor fix. | 2020-10-24T04:31:21 | 2020-10-24T04:31:21 | {} | MEMBER |
ClickHouse | ClickHouse | 715,672,131 | 16,277 | alexey-milovidov | @filipecaixeta We are slightly changed the interface of IFunction in the meantime (for better).
Some adaptation is needed to build code after these changes. | 2020-10-24T04:35:17 | 2020-10-24T04:35:17 | {} | MEMBER |
ClickHouse | ClickHouse | 715,854,991 | 10,540 | alexey-milovidov | @ucasFL Great, it's yours :+1: | 2020-10-24T07:26:01 | 2020-10-24T07:26:01 | {} | MEMBER |
ClickHouse | ClickHouse | 715,936,254 | 16,336 | den-crane | You use obsolete version 20.3.5.21. It's better to upgrade CH to 20.3.20.6.
>clickhouse-client --user default --password super --database system --multiline --query="show databases"
>Code: 210. DB::NetException: Connection reset by peer, while reading from socket (127.0.0.1:9000)
>curl 'http://172.16.24.133:... | 2020-10-24T15:27:46 | 2020-10-24T15:27:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 716,041,760 | 8,828 | fibersel | @alexey-milovidov As I have figured out, ```clickhouse-client``` doesn't unpack compressed file itself, it sends data to server to be unpacked there.So the problem is to add ```Content-Encoding: zstd``` option? | 2020-10-24T19:14:03 | 2020-10-24T21:06:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 716,084,675 | 10,540 | alexey-milovidov | Subqueries can be grouped by parentheses. The logical order is defined by these parentheses.
If there are no parentheses, we should assume that UNION operation is left-associative.
E.g. if we have a query:
```
SELECT 1
UNION ALL
SELECT 1
UNION
SELECT 2
UNION ALL
SELECT 2
```
it is the same as
```
... | 2020-10-25T02:41:19 | 2020-10-25T02:45:36 | {} | MEMBER |
ClickHouse | ClickHouse | 715,362,354 | 15,934 | tavplubix | Yandex third-party checks failure is related | 2020-10-23T14:03:18 | 2020-10-23T14:03:18 | {} | MEMBER |
ClickHouse | ClickHouse | 715,582,407 | 16,307 | den-crane | related https://github.com/ClickHouse/ClickHouse/issues/11933#issuecomment-661310032 (huge number or parts) | 2020-10-23T20:47:32 | 2020-10-23T20:48:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 715,650,569 | 16,265 | bohutang | 3 is possible.
But the check tool should extends and dump the event datas.
Any way to give your binlog file? | 2020-10-24T01:19:10 | 2020-10-24T01:19:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 715,661,533 | 16,008 | zhang2014 | reference: https://github.com/mysql/mysql-server/blob/8.0/sql/sql_yacc.yy#L7131 | 2020-10-24T02:49:51 | 2020-10-24T02:49:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 715,668,716 | 16,330 | alexey-milovidov | It can be optimized if we do OR of null masks before call to aggregate function - it's worth doing. | 2020-10-24T03:55:37 | 2020-10-24T03:55:37 | {} | MEMBER |
ClickHouse | ClickHouse | 715,878,309 | 16,249 | alexey-milovidov | @vzakaznikov It didn't help: https://clickhouse-test-reports.s3.yandex.net/16332/96b932f2291368fc608a38fd24c60fe31f0e5fd1/testflows_check.html#fail1 | 2020-10-24T08:08:37 | 2020-10-24T08:08:37 | {} | MEMBER |
ClickHouse | ClickHouse | 715,908,026 | 16,121 | azat | So now looks OK (although there is minimal conflict in the fasttest skip list, it is pretty trivial so I will not solve it by myself to avoid trigerring CI) | 2020-10-24T12:29:08 | 2020-10-24T12:29:08 | {} | MEMBER |
ClickHouse | ClickHouse | 716,086,362 | 16,236 | trollhe | this table is large,about 200 billions rows. about 3 tb gz file.this one day data size .
Can you help me？ | 2020-10-25T03:03:30 | 2020-10-25T03:04:13 | {} | NONE |
ClickHouse | ClickHouse | 716,128,188 | 16,349 | sundy-li | Is that any need for ClickHouse to integration with hive metastore? | 2020-10-25T11:00:00 | 2020-10-25T11:00:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 715,433,705 | 16,261 | alesapin | `01516_drop_table_stress` -- timeout https://clickhouse-test-reports.s3.yandex.net/16261/b29cfa7f826566192478c77413ccca34e8a719ac/functional_stateless_tests_(memory).html#fail1 | 2020-10-23T16:06:12 | 2020-10-23T16:17:21 | {} | MEMBER |
ClickHouse | ClickHouse | 715,501,134 | 15,144 | azat | @akuzm just out of curiosity, what does it improves in perf tests? (taking into account that fact that you disable percpu arena, so you may have a little bit higher memory reserved, and sometimes a little bit less overhead on freeing, but this just a possibility) | 2020-10-23T18:20:28 | 2020-10-23T18:20:28 | {} | MEMBER |
ClickHouse | ClickHouse | 715,522,660 | 16,127 | tavplubix | AST fuzzer - #15889
Performance - unrelated
Yandex synchronization check - https://nda.ya.ru/t/cyaxqbD_3ZQjkt | 2020-10-23T18:52:59 | 2020-10-23T18:52:59 | {} | MEMBER |
ClickHouse | ClickHouse | 715,631,680 | 16,121 | azat | So the problem was only in the test
Due to metric_log it has some difference, of course I could rewrite the test to allow some small delta, ~20K is enough, but I want ensure/show that after some actions the MemoryTracking does not changes at all.
So the stateless test was replaced with integration test (basically i... | 2020-10-23T23:31:00 | 2020-10-23T23:31:00 | {} | MEMBER |
ClickHouse | ClickHouse | 715,671,561 | 16,265 | liangli1213 | @BohuTANG 'But the check tool should extends' - What does extend here mean?
I can't give you the binlog since it is too big as well as has secrete data, if possible, you can tell me what you want to do with the binlog, and I can send you the abnormal piece of data.
-Lyon | 2020-10-24T04:28:40 | 2020-10-24T04:28:40 | {} | NONE |
ClickHouse | ClickHouse | 715,705,411 | 16,265 | zhang2014 | @liangli1213 Maybe you could provide your binlog file? | 2020-10-24T05:11:21 | 2020-10-24T05:11:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 715,910,364 | 16,249 | vzakaznikov | > @vzakaznikov It didn't help: https://clickhouse-test-reports.s3.yandex.net/16332/96b932f2291368fc608a38fd24c60fe31f0e5fd1/testflows_check.html#fail1
Investigating. It is a different fail `/clickhouse/ldap/external user directory/connection protocols/plain text`. | 2020-10-24T12:49:50 | 2020-10-24T12:49:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 716,082,922 | 10,540 | ucasfl | Hi, I have some question about this task, do we support such as `SELECT ``` UNION SELECT ... UNION ALL/UNION DISTINCT SELECT ...`, and if support, we should compute in what order? @alexey-milovidov | 2020-10-25T02:18:45 | 2020-10-25T02:18:45 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 716,100,893 | 16,236 | amosbird | zlib inflate will process as much input as possible using the available output. Once it can neither process input nor produce output, it will return buffer error. So this looks like the input stream is ended unexpectedly. Perhaps something is flaky inside libhdfs3. You may try using native hdfs client to read the file ... | 2020-10-25T06:23:07 | 2020-10-25T06:23:07 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 715,397,810 | 16,251 | keliss | Oh, sorry, I calculated incorrectly. Is there any particular reason to keep some limit for DROP queries at all? I can't imagine a use-case for this setting. | 2020-10-23T15:03:25 | 2020-10-23T15:03:25 | {} | NONE |
ClickHouse | ClickHouse | 715,922,823 | 12,632 | den-crane | It's not true and documentation is incorrect.
` alter table sp modify setting storage_policy='default1'`
New policy should include all disks / volumes which were included into old one.
After that you can move all tables to a new volume. And remove old volumes from the new policy. | 2020-10-24T14:27:41 | 2020-10-24T14:27:41 | {
"+1": 12
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 715,957,165 | 16,279 | den-crane | Feature: Filter (prune) hdfs files by condition in `select from hdfs(*) where _file = ` but not by globe/wildcard in hfds(....). | 2020-10-24T16:01:51 | 2020-10-24T16:01:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 716,075,542 | 16,328 | l1t1 | any datasheet of the performance improvements? | 2020-10-25T00:50:24 | 2020-10-25T00:50:24 | {} | NONE |
ClickHouse | ClickHouse | 715,381,288 | 16,284 | alesapin | > Perhaps we can serialize the metadata with `ORDER BY` clause generated?
Yes, but it can be surprising for the user. You just typed one `CREATE` query, and `SHOW CREATE TABLE` shows you different one. | 2020-10-23T14:35:51 | 2020-10-23T14:35:51 | {} | MEMBER |
ClickHouse | ClickHouse | 715,418,033 | 15,934 | tavplubix | Now Yandex third-party checks are ok. | 2020-10-23T15:37:38 | 2020-10-23T15:37:38 | {} | MEMBER |
ClickHouse | ClickHouse | 715,419,365 | 16,251 | abyss7 | > Oh, sorry, I calculated incorrectly. Is there any particular reason to keep some limit for DROP queries at all? I can't imagine a use-case for this setting.
From documentation:
> In many cases mistakes like these will affect all replicas. ClickHouse has built-in safeguards to prevent some types of mistakes — for ... | 2020-10-23T15:40:05 | 2020-10-23T15:40:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 715,878,485 | 16,044 | alexey-milovidov | @vzakaznikov @MyroTk
test_distributed_over_live_view failed.
TestFlows LDAP failed. | 2020-10-24T08:10:31 | 2020-10-24T08:10:31 | {} | MEMBER |
ClickHouse | ClickHouse | 716,036,960 | 15,801 | alexey-milovidov | I'm about \u escape seqences - looks like some automated change, it makes text less readable. | 2020-10-24T18:38:04 | 2020-10-24T18:38:04 | {} | MEMBER |
ClickHouse | ClickHouse | 715,378,795 | 16,276 | den-crane | `max_memory_usage -- Maximum memory usage for processing of single query. Zero means unlimited.`
Two parallel session can use 2*max_memory_usage.
`max_memory_usage_for_all_queries -- Maximum memory usage for processing all concurrently running queries on the server. Zero means unlimited.`
CH usage memory = `s... | 2020-10-23T14:31:17 | 2020-10-23T14:31:17 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 715,668,810 | 16,330 | alexey-milovidov | The case of aggregation without GROUP BY keys can also be greatly optimized (by pre-filtering by null masks). | 2020-10-24T03:56:41 | 2020-10-24T03:56:41 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 715,910,877 | 16,306 | filimonov | May be we need to split total bytes into total_bytes_on_disk / total_bytes_in_ram? | 2020-10-24T12:54:16 | 2020-10-24T12:54:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 716,073,930 | 16,345 | l1t1 | the URL is banned in my area | 2020-10-25T00:31:14 | 2020-10-25T00:31:14 | {} | NONE |
ClickHouse | ClickHouse | 716,076,492 | 16,328 | alexey-milovidov | https://clickhouse-test-reports.s3.yandex.net/15239/677a7e485f511314cd7bfbd349cf6d24ccb2be37/performance_comparison/report.html#fail1
(from here: https://github.com/ClickHouse/ClickHouse/pull/15239) | 2020-10-25T00:59:24 | 2020-10-25T00:59:24 | {} | MEMBER |
ClickHouse | ClickHouse | 721,406,944 | 16,637 | alexey-milovidov | If it's performance improvement, why "bug fix"? | 2020-11-03T22:27:47 | 2020-11-03T22:27:47 | {} | MEMBER |
ClickHouse | ClickHouse | 721,419,131 | 8,772 | alexey-milovidov | > If I change the order by key to be `os`
That's expected because data will become sorted and compress better. | 2020-11-03T23:01:47 | 2020-11-03T23:01:47 | {} | MEMBER |
ClickHouse | ClickHouse | 721,493,748 | 16,655 | Slach | @Enmk @filimonov please look this issue | 2020-11-04T03:37:38 | 2020-11-04T03:37:38 | {} | CONTRIBUTOR |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.